Background:
As subtypes of drug-induced akathisia have become accepted and attempts have been made at establishing diagnostic criteria, a prospective study of the clinical features and predisposing factors of acute akathisia is a significant deficiency in the literature.
Methods:
One hundred consecutive inpatients with nonorganic psychotic disorders, not receiving neuroleptics or other drugs and free of akathisia and related disorders at admission, were assessed for psychiatric status and movement disorders at baseline and daily for 2 weeks, with detailed examinations on days 7 and 14. Multiple operational criteria for akathisia were used. The following risk factors were examined: age, sex, current neuroleptic dose, rate of increment of dose, drug type, duration of illness, past use of neuroleptics, extrapyramidal side effects score, Zung Depression Scale score, Spielberger State Anxiety Inventory score, psychosis score, and smoking.
Results:
Using a global rating, mild akathisia developed in 41% and moderate-to-severe akathisia in 21%. The symptoms that best discriminated akathisia from nonakathisia were shifting weight from foot to foot or walking on the spot, inability to keep legs still, feelings of inner restlessness, and shifting of body position in the chair. The subjective and objective symptoms loaded on separate factors. Akathisia ratings had low correlations with the anxiety and Zung scores. Receiver operating characteristic analysis suggested a cutoff score of 4 on our 10-item Akathisia Scale as optimal for the diagnosis of akathisia, with a stricter criterion of scores of 2 or more on both the subjective and objective items being more suitable for research diagnosis. The most significant predisposing factors were the extrapyramidal side effects score and current neuroleptic dose and its rate of increment, with lesser contributions from serum iron status and medication type. Predictability was, however, modest.
Conclusion:
Acute akathisia is a common syndrome with well-defined clinical features. Its occurrence can be predicted with only modest accuracy.